Abstract
Fifty jejunal biopsy specimens obtained from normal subjects and from untreated and treated patients with coeliac disease were assessed blindly by three independent observers, each of them using different morphometric techniques-namely, linear measurement, stereology, and computer aided microscopy. In two of 26 control biopsy specimens linear measurement was not possible because of distortion of villi. Highly significant (p less than 0.001) correlation coefficients were found between the different techniques. With all methods significant differences between controls and patients with coeliac disease and between treated and untreated coeliac patients were found. Only by stereology, however, was there no overlap between results for patients and those for controls. In view of the limitations of linear measurement and the high cost and complexity of computer aided microscopy, we propose that a simple stereological technique using an eyepiece graticule is the method of choice in the quantitative assessment of mucosal architecture in jejunal biopsy specimens.